
GWAF2.0 version made the following changes from previous versions.

1. Including scripts for genome-wide interaction analyses with family data.

2. Replacing gee package with geepack package in scripts for binary traits.

3. Including GEE as an additional option for analyzing continuous traits.

4. Correcting reading in gzipped genotype file to un-gzipped genotype file (by removing gzfile in read.in.data function in scripts)

5. Correcting a separator bug in read.in.data function in scripts to allow reading in input files using other separator format properly. Thanks, Xiang Li! 
Previuos: read.in.data <- function(phenfile,genfile,pedfile,sep.ped=",",sep.phe=",",sep.gen=",") 
Now: read.in.data <- function(phenfile,genfile,pedfile,sep.ped=sep.ped,sep.phe=sep.phe,sep.gen=sep.gen) 

6. Adding class(lme.out)!="try-error" checking codition in LME scripts, so when false, NA result will be output and the analysis will not be terminated with the following error. 
   Thanks, Yan Meng.
>>Error in gchol.bdsmatrix(x, tolerance) : NA/NaN/Inf in foreign function call (arg 4)  
>>Error in sqrt(lme.out$var[2, 2]) : Non-numeric argument to mathematical function

7. Correction made to manual: genotype coding is based on the (expected) number of coded alleles, not limited to minor alleles. Thanks, Rui Li.

8. Correction made to fix gee.lgst() that reports missing genotype proportions of 0.5 when there are no missing data. Thanks, Thomas Lumley.

9. Correction made to auto() function that deleted all generated R/shell scripts (*.R/sh) when a submitted job is complete. Now when a submitted job is complete, only the R/shell
scripts of the submitted job will be deleted.  

10. Change dependence on kinship package that has been archived by including functions from kinship package.


GWAF2.1 version made the following changes from previous versions.

1. Fixing subdirectory issue (GWAF/GWAF/) when installing GWAF

2. Removing functions/scripts from archived kinship package and now depending on coxme package, necessary changes in LME scripts/functions are made.

2. Removing genotype count/MAF filter(s) in LME functions and geepack.quant functions. lmepack.batch is recommended for analyzing single rare variants with continuous traits.

3. Including GLMM for analyzing binary traits with observed genotype data. Recommended for analyzing single rare variants. 

4. Including lmeVpack.batch.imputed function, a more efficient and faster version of lmepack.batch.imputed, for analyzing continuous traits with imputed genotype data.










